[issue45863] tarfile zeroes ustar header fields unnecessarily
Joshua Root added the comment: PR has been marked stale; friendly ping. -- ___ Python tracker <https://bugs.python.org/issue45863> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue45863] tarfile zeroes ustar header fields unnecessarily
Change by Joshua Root : -- keywords: +patch pull_requests: +27931 stage: -> patch review pull_request: https://github.com/python/cpython/pull/29693 ___ Python tracker <https://bugs.python.org/issue45863> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue45863] tarfile zeroes ustar header fields unnecessarily
New submission from Joshua Root : When using the pax format, tarfile will zero out the field in the ustar header for any values that are represented with a float, setting the correct value only in the pax header. This mainly seems to apply to the mtime. This behaviour doesn't cause problems when using an unarchiver that understands the pax header, but unarchivers that don't will extract incorrect metadata (most obviously all mtimes set to the epoch). Compatibility with such unarchivers can easily be achieved by rounding the float value to int for the ustar header only, thus at least giving mtimes that are accurate to the nearest second instead of nothing. -- components: Library (Lib) messages: 406744 nosy: jmr priority: normal severity: normal status: open title: tarfile zeroes ustar header fields unnecessarily type: behavior versions: Python 3.10, Python 3.11, Python 3.9 ___ Python tracker <https://bugs.python.org/issue45863> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue44853] 3.10.0rc1 published md5 and size do not match for source archives
Change by Joshua Root : -- title: 3.1.10rc1 published md5 and size do not match for source archives -> 3.10.0rc1 published md5 and size do not match for source archives ___ Python tracker <https://bugs.python.org/issue44853> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue44853] 3.1.10rc1 published md5 and size do not match for source archives
New submission from Joshua Root : The download page lists the following: Filenamemd5 size Python-3.10.0rc1.tgzc051bf7a52a45cb0ec2cefbe915417e1 40764776 Python-3.10.0rc1.tar.xz 2861cdd4cf71c6425fde1fedc14bb283 28197832 The downloaded files instead have these properties: Python-3.10.0rc1.tgzd23c2a8228705b17e8414f1660e4bb73 24955561 Python-3.10.0rc1.tar.xz edd2eb2f7f4a932ed59196cbe373e5fb 18680452 The gpg signatures do verify ok however. The md5 and size listed for the macOS installer seem to be correct. I didn't check the Windows installers. -- components: Installation messages: 399103 nosy: jmr, pablogsal priority: normal severity: normal status: open title: 3.1.10rc1 published md5 and size do not match for source archives type: security versions: Python 3.10 ___ Python tracker <https://bugs.python.org/issue44853> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue43568] Drop support for Mac OS X < 10.3 module linking
Change by Joshua Root : -- pull_requests: +24654 pull_request: https://github.com/python/cpython/pull/26001 ___ Python tracker <https://bugs.python.org/issue43568> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue43568] Drop support for Mac OS X < 10.3 module linking
Joshua Root added the comment: The part that is a bug is that the whole version was checked, not just the major version--you couldn't target macOS 11.0 if your Python was built for 11.3, for example. MacPorts users getting an error in that situation was the original motivation for coming up with the distutils change. -- ___ Python tracker <https://bugs.python.org/issue43568> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue43568] Drop support for Mac OS X < 10.3 module linking
Joshua Root added the comment: Thanks Ned. Would it be OK to backport just the distutils change to 3.9? -- ___ Python tracker <https://bugs.python.org/issue43568> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue43568] Drop support for Mac OS X < 10.3 module linking
Joshua Root added the comment: Here's the PR for pypa/distutils, which is just relaxing the MDT check since it has to work with older Python versions: https://github.com/pypa/distutils/pull/36 -- ___ Python tracker <https://bugs.python.org/issue43568> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue43568] Drop support for Mac OS X < 10.3 module linking
Change by Joshua Root : -- keywords: +patch pull_requests: +23702 stage: -> patch review pull_request: https://github.com/python/cpython/pull/24941 ___ Python tracker <https://bugs.python.org/issue43568> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue43568] Drop support for Mac OS X < 10.3 module linking
New submission from Joshua Root : The `-undefined dynamic_lookup` option can only be used in LDSHARED on Mac OS X 10.3 and later. There is a fallback to explicitly linking with the framework for 10.2 and earlier. I'm pretty sure that currently supported Python versions don't build on 10.2 or older for several other reasons (I happen to know that even building on 10.5 requires a little patching.) So it's probably reasonable to just drop this code path. There is a closely related check in distutils, though you would only know it's related if you looked through the history as I did. It errors out if you try to build a module for an older MACOSX_DEPLOYMENT_TARGET than Python was configured with. The purpose of that is to prevent using the wrong LDSHARED flags for the target platform. If 10.2 support is dropped, that check can be removed entirely. I am aware that distutils is deprecated, going away, etc., and I am submitting a PR to setuptools as well. But setuptools does not yet override the stdlib distutils with its own by default, so bugs in the stdlib copy are still relevant. If it's decided to keep 10.2 support, the check in distutils should still be relaxed to error only if the current MDT is < 10.3 and the configured MDT is >= 10.3. I can easily put together a PR for that if needed. Either way, the approach taken in setuptools will depend on how LDSHARED is handled here. -- components: Build, Distutils, macOS messages: 389157 nosy: dstufft, eric.araujo, jmr, ned.deily, ronaldoussoren priority: normal severity: normal status: open title: Drop support for Mac OS X < 10.3 module linking type: enhancement versions: Python 3.10, Python 3.6, Python 3.7, Python 3.8, Python 3.9 ___ Python tracker <https://bugs.python.org/issue43568> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue42692] Build fails on macOS when compiler doesn't define __has_builtin
Change by Joshua Root : -- keywords: +patch pull_requests: +22735 stage: -> patch review pull_request: https://github.com/python/cpython/pull/23873 ___ Python tracker <https://bugs.python.org/issue42692> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue42692] Build fails on macOS when compiler doesn't define __has_builtin
New submission from Joshua Root : The line in posixmodule.c that checks for __builtin_available is rejected by compilers that don't have __has_builtin. The second check needs to be in a nested #if. -- components: Build, macOS messages: 383437 nosy: jmr, ned.deily, ronaldoussoren priority: normal severity: normal status: open title: Build fails on macOS when compiler doesn't define __has_builtin type: compile error versions: Python 3.10, Python 3.9 ___ Python tracker <https://bugs.python.org/issue42692> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue41100] Support macOS 11 and Apple Silicon Macs
Joshua Root added the comment: Are there plans to backport PR 22855 to any branches older than 3.9? -- nosy: +jmr ___ Python tracker <https://bugs.python.org/issue41100> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue41617] __builtin_bswap16 is used without checking it is supported
Joshua Root added the comment: Confirmed fixed in 3.9.1 and 3.10.0a3. Thanks Victor! -- ___ Python tracker <https://bugs.python.org/issue41617> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue42598] Some configure checks rely on implicit function declaration
Change by Joshua Root : -- keywords: +patch pull_requests: +22558 stage: -> patch review pull_request: https://github.com/python/cpython/pull/23690 ___ Python tracker <https://bugs.python.org/issue42598> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue42598] Some configure checks rely on implicit function declaration
New submission from Joshua Root : There are several cases in the configure script of exit being used without including stdlib.h, and one case of close being used without including unistd.h. A compiler that does not allow implicit function declaration, such as clang in Xcode 12, will error out due to this, which may cause the affected checks to give incorrect results. -- components: Build messages: 382726 nosy: jmr priority: normal severity: normal status: open title: Some configure checks rely on implicit function declaration type: compile error versions: Python 3.10, Python 3.9 ___ Python tracker <https://bugs.python.org/issue42598> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue31601] Availability of utimensat, futimens not checked correctly on macOS
Joshua Root added the comment: Looks like this is fixed as of 3.9.1. -- resolution: -> fixed stage: -> resolved status: open -> closed ___ Python tracker <https://bugs.python.org/issue31601> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue41617] __builtin_bswap16 is used without checking it is supported
Joshua Root added the comment: > I'm curious. Can I ask you on which platform do you use clang older than 3.2? Mac OS X 10.7 / Xcode 4.6.3. I'm not using it personally, but we have automated builds on that platform. Unfortunately the patch ultimately committed did not fix the build there. Clang reports its version as "Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn)". __clang_major__ is 4 and __clang_minor__ is 2. Apple's versioning scheme is different to that of LLVM upstream, which is one reason why I preferred detecting features directly rather than inserting externally-derived knowledge about which versions provide which features. Apologies for not getting back to you about this sooner; the notifications appear to have gotten lost. -- status: closed -> open ___ Python tracker <https://bugs.python.org/issue41617> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue41617] __builtin_bswap16 is used without checking it is supported
Change by Joshua Root : -- pull_requests: +21053 pull_request: https://github.com/python/cpython/pull/21943 ___ Python tracker <https://bugs.python.org/issue41617> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue41617] __builtin_bswap16 is used without checking it is supported
Change by Joshua Root : -- versions: +Python 3.10 ___ Python tracker <https://bugs.python.org/issue41617> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue41617] __builtin_bswap16 is used without checking it is supported
Change by Joshua Root : -- keywords: +patch pull_requests: +21052 stage: -> patch review pull_request: https://github.com/python/cpython/pull/21942 ___ Python tracker <https://bugs.python.org/issue41617> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue41617] __builtin_bswap16 is used without checking it is supported
New submission from Joshua Root : Older clang versions don't have __builtin_bswap16, but it's always used when compiling with clang, which means the build fails with those older versions. The code should use __has_builtin to check. -- components: Build messages: 375806 nosy: jmr priority: normal severity: normal status: open title: __builtin_bswap16 is used without checking it is supported type: compile error versions: Python 3.9 ___ Python tracker <https://bugs.python.org/issue41617> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue38360] single-argument form of -isysroot should be supported
Joshua Root added the comment: That ValueError I mentioned causes build failures for extension modules whenever the CFLAGS in sysconfig contains an -isysroot flag in the single arg form. We ran into it a lot in MacPorts on Mojave and Catalina. So I would consider it a bug, and would prefer to backport to all branches that are open for bug fixes. -- ___ Python tracker <https://bugs.python.org/issue38360> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue38360] single-argument form of -isysroot should be supported
Joshua Root added the comment: Ping? -- ___ Python tracker <https://bugs.python.org/issue38360> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue38014] ./configure on recent macOS versions can fail with multiple "present but cannot be compiled" messages
Joshua Root added the comment: The crash reported by BMinas is reputedly related to the -fstack-check option, which is said to be on by default when MACOSX_DEPLOYMENT_TARGET=10.15. I've been unable to confirm whether there was just a bug in the beta, or if it's still a problem on the final Catalina release. -- nosy: +jmr ___ Python tracker <https://bugs.python.org/issue38014> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue38360] single-argument form of -isysroot should be supported
New submission from Joshua Root : The path associated with the -isysroot compiler flag can be supplied either as a separate argument or in the same argument as -isysroot itself. The places in library code that do special handling of this flag should support both forms, but currently only support the two separate arguments form. This means that the flag may not be removed when pointing to a nonexistent SDK or when a different SDK is specified in the user's CFLAGS, and at worst a ValueError is raised in compiler_fixup. -- components: Distutils, macOS messages: 353838 nosy: dstufft, eric.araujo, jmr, ned.deily, ronaldoussoren priority: normal pull_requests: 16146 severity: normal status: open title: single-argument form of -isysroot should be supported type: behavior versions: Python 2.7, Python 3.5, Python 3.6, Python 3.7, Python 3.8, Python 3.9 ___ Python tracker <https://bugs.python.org/issue38360> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue34652] never enable lchmod on Linux
Joshua Root added the comment: The follow-up fix (AC_CHECK_FUNC -> AC_CHECK_FUNCS) still needs to be backported to 2.7. Currently the lack of it causes lchmod to be disabled on all platforms, not just Linux. -- nosy: +jmr ___ Python tracker <https://bugs.python.org/issue34652> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue31601] Availability of utimensat, futimens not checked correctly on macOS
Joshua Root added the comment: This isn't just a cross-build issue, or rather (depending on how you look at it) cross-builds are increasingly the norm due to Apple only providing a newer SDK on older OS versions. For example the latest version of Xcode available for 10.12 only comes with the 10.13 SDK. -- ___ Python tracker <https://bugs.python.org/issue31601> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue31601] Availability of utimensat and futimens not checked correctly on macOS
New submission from Joshua Root : The futimens and utimensat functions are available in High Sierra but not previous versions of macOS. Building against the 10.13 SDK and deploying to an earlier version will result in a crash when the dynamic linker can't find the symbols at runtime. Availability of the symbols needs to be checked at runtime before use. This was originally reported in <https://trac.macports.org/ticket/54893>. -- components: Build messages: 303098 nosy: jmr priority: normal severity: normal status: open title: Availability of utimensat and futimens not checked correctly on macOS type: crash versions: Python 3.5, Python 3.6 ___ Python tracker <https://bugs.python.org/issue31601> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue11149] [PATCH] Configure should enable -fwrapv for clang
Joshua Root added the comment: The fix that was committed doesn't work if CC is a full path like /usr/bin/clang. -- nosy: +jmr ___ Python tracker <http://bugs.python.org/issue11149> ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue6802] build fails on Snow Leopard
Joshua Root added the comment: It works now. Thanks! -- ___ Python tracker <http://bugs.python.org/issue6802> ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue6802] build fails on Snow Leopard
Joshua Root added the comment: Should line 110 of Lib/plat-mac/macresource.py say 'isinstance' rather than 'instance'? -- ___ Python tracker <http://bugs.python.org/issue6802> ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue6802] build fails on Snow Leopard
Joshua Root added the comment: Tried again after ensuring that ~/test was completely empty; no difference. -- ___ Python tracker <http://bugs.python.org/issue6802> ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue6802] build fails on Snow Leopard
Joshua Root added the comment: I tried release26-maint just now (r74683) and it errored out during make install. New log attached. -- Added file: http://bugs.python.org/file14846/release26-maint.log ___ Python tracker <http://bugs.python.org/issue6802> ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue6802] build fails on Snow Leopard
Joshua Root added the comment: Noting for completeness that: * 3.0.1 behaves the same as 3.1.1 (works without --enable-framework). * 2.5.4 fails with 'cc1: error: unrecognized command line option "-Wno- long-double"'. Once this is removed from the configure script, it then fails in mactoolboxglue.c due to missing Carbon functions, whether or not it was configured with --enable-framework. Using --disable-toolbox- glue and not using --enable-framework allows the build to succeed. Using --disable-toolbox-glue and --enable-framework causes a failure involving struct rusage in posixmodule.c. Adding "#define _DARWIN_C_SOURCE" to pyconfig.h fixes this, but then there's the same "file is not of required architecture" failure seen in later versions. * 2.4.6 behaves the same as 2.5.4 apart from not needing the _DARWIN_C_SOURCE addition. * None of the failures are affected by /opt/local/bin's presence or absence in my PATH. -- versions: +Python 2.4, Python 2.5, Python 3.0 ___ Python tracker <http://bugs.python.org/issue6802> ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue6802] build fails on Snow Leopard
Joshua Root added the comment: Note that this problem appears to be x86_64-specific (caused by "- arch_only i386"), and only happens if you configure with --enable- framework. -- ___ Python tracker <http://bugs.python.org/issue6802> ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue6802] build fails on Snow Leopard
New submission from Joshua Root : Building Python 2.6.2 on Mac OS X 10.6 (final version, 10A432) fails. Full build transcript is attached. Other system details: Xcode 3.2, hardware is MacPro1,1 -- assignee: ronaldoussoren components: Build, Macintosh files: py26-snowleopard.log messages: 92067 nosy: jmr, ronaldoussoren severity: normal status: open title: build fails on Snow Leopard type: compile error versions: Python 2.6 Added file: http://bugs.python.org/file14801/py26-snowleopard.log ___ Python tracker <http://bugs.python.org/issue6802> ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com